基于改进的KPCA-PSO-WLSSVM在循环水腐蚀预测中的应用
摘要
针对石化现场腐蚀速率参数存在测量成本高、测量周期长的问题,结合国内外腐蚀研究现状提出了一种基于改进的自适应加权最小二乘支持向量机回归建模方法。该方法首先对数据进行整合处理,借助核主成分分析(KPCA)算法对整合后的数据进行主成分提取,依据处理好的数据建立LS-SVM模型;其次采用改进的加权算法对LS-SVM进行权值处理;然后采用全局搜索能力较强的混沌粒子群-模拟退火优化算法(CPSO-SA)对LS-SVM模型正则化参数和核宽度参数进行优化,提高模型的泛化能力;最后建立优化后的KPCA-WLS-SVM模型。实验结果表明,应用该方法建立的循环水腐蚀预测模型的预测准确度远远高于其他预测模型的预测准确度。
引文
[1]喻西崇,赵金洲,吴应湘,等.利用改进的神经网络预测腐蚀管道的剩余强度[J].压力容器,2003(10):22-27.
[2]李亚峰,陶翠翠,蒋白懿,等.改进BP神经网络预测埋地金属给水管道外腐蚀速率的研究[J].给水排水,2010,46(增刊1):440-442.
[3]董超,胡艳珍,李晨光.基于改进的T-S模糊神经网络循环冷却水腐蚀预测[J].化工自动化及仪表,2018,45(1):51-55.
[4]战越,宋志刚,王海莹.用SVM预测稀硫酸对透水混凝土的腐蚀[J].价值工程,2017,36(1):91-92.
[5]骆正山,毕傲睿,王小完.基于PCA-SVM的高含硫油气混输管道腐蚀预测[J].中国安全科学学报,2016,26(2):85-90.
[6]何敏.炼油装置循环水换热系统腐蚀模型与评价研究[D].西安:西安石油大学,2014.
[7]Alireza Baghban,Mohammad Navid Kardani,Sajjad Habibzadeh.Prediction viscosity of ionic liquids using a hybrid LSSVM and group contribution method[J].Journal of Molecular Liquids,2017,236:452-464.
[8]Hamidreza Yarveicy,Mohammad M Ghiasi.Modeling of gas hydrate phase equilibria:Extremely randomized trees and LSSVM approaches[J].Journal of Molecular Liquids,2017,243:533-541.
[9]龙文,焦建军,龙祖强.基于PSO优化LSSVM的未知模型混沌系统控制[J].物理学报,2011,60(11):120-125.
[10]Suykens J A K,Vandewalle J.Least squares support vector machines classifiers[J].Neural Processing Letters,1999,9(3):293-300.
[11]Sheng Hanlin,Zhang Tianhong.Aircraft engine thrust estimator design based on GSA-LSSVM[J].International Journal of Turbo&Jet-Engines,2017,34(3).
[12]Ma Liang,Liu Xinggao.A novel APSO-aided weighted LSSVM method for nonlinear hammerstein system identification[J].Journal of the Franklin Institute,2017,354(4):1892-1906.
[13]Gorjaei R G,Songolzadeh R,Torkaman M,et al.A novel PSO-LSSVM model for predicting liquid rate of two phase flow through wellhead chokes[J].Journal of Natural Gas Science and Engineering,2015,24:228-237.
[14]Gao Junwei,Leng Ziwen,Qin Yong,et al.Application of GA-LSSVM in fault diagnosis of subway auxiliary inverter[M].Berlin Heidelberg:Springer,2014.
[2]李亚峰,陶翠翠,蒋白懿,等.改进BP神经网络预测埋地金属给水管道外腐蚀速率的研究[J].给水排水,2010,46(增刊1):440-442.
[3]董超,胡艳珍,李晨光.基于改进的T-S模糊神经网络循环冷却水腐蚀预测[J].化工自动化及仪表,2018,45(1):51-55.
[4]战越,宋志刚,王海莹.用SVM预测稀硫酸对透水混凝土的腐蚀[J].价值工程,2017,36(1):91-92.
[5]骆正山,毕傲睿,王小完.基于PCA-SVM的高含硫油气混输管道腐蚀预测[J].中国安全科学学报,2016,26(2):85-90.
[6]何敏.炼油装置循环水换热系统腐蚀模型与评价研究[D].西安:西安石油大学,2014.
[7]Alireza Baghban,Mohammad Navid Kardani,Sajjad Habibzadeh.Prediction viscosity of ionic liquids using a hybrid LSSVM and group contribution method[J].Journal of Molecular Liquids,2017,236:452-464.
[8]Hamidreza Yarveicy,Mohammad M Ghiasi.Modeling of gas hydrate phase equilibria:Extremely randomized trees and LSSVM approaches[J].Journal of Molecular Liquids,2017,243:533-541.
[9]龙文,焦建军,龙祖强.基于PSO优化LSSVM的未知模型混沌系统控制[J].物理学报,2011,60(11):120-125.
[10]Suykens J A K,Vandewalle J.Least squares support vector machines classifiers[J].Neural Processing Letters,1999,9(3):293-300.
[11]Sheng Hanlin,Zhang Tianhong.Aircraft engine thrust estimator design based on GSA-LSSVM[J].International Journal of Turbo&Jet-Engines,2017,34(3).
[12]Ma Liang,Liu Xinggao.A novel APSO-aided weighted LSSVM method for nonlinear hammerstein system identification[J].Journal of the Franklin Institute,2017,354(4):1892-1906.
[13]Gorjaei R G,Songolzadeh R,Torkaman M,et al.A novel PSO-LSSVM model for predicting liquid rate of two phase flow through wellhead chokes[J].Journal of Natural Gas Science and Engineering,2015,24:228-237.
[14]Gao Junwei,Leng Ziwen,Qin Yong,et al.Application of GA-LSSVM in fault diagnosis of subway auxiliary inverter[M].Berlin Heidelberg:Springer,2014.